Before NFTs Had a Name
The history of NFTs begins with a simple problem: What would you do if you wanted to collect something that did not physically exist? You could save a picture to your computer, of course. But saving a picture is not quite the same as owning it. Anyone else could save exactly the same picture. There was no obvious way to tell which copy was the original, who owned it, or whether the artist had intended the file to be a collectible in the first place.
In January 1993, the cryptographer Hal Finney wrote about an idea he called “crypto trading cards”: digital collectibles that could be bought, sold and traded using cryptography. It was an intriguing idea. The machinery needed to make it practical did not yet exist.
Sixteen years later, Bitcoin would provide part of that machinery.
Bitcoin changes the question
When Bitcoin appeared in 2009, its purpose was fairly specific: create digital money that could be transferred without a bank keeping the master record. The clever part was not simply that Bitcoin was digital. Plenty of things were digital already. The clever part was keeping a shared record of transactions so that everyone could agree which coins had been spent and which had not. This solved the double-spending problem. If I send you a Bitcoin, I cannot simply send the same Bitcoin to somebody else and expect both transactions to be accepted. The network keeps track of what has happened.
Once people had this new kind of public ledger, it was natural to wonder whether it could keep track of other things as well. What if the thing being recorded wasn't money? What if it was a share, a domain name, a piece of artwork or a collectible?
That question led to some of the earliest experiments in digital collectibles.
Colouring coins
One of the first ideas was Colored Coins, proposed in 2012. The name sounds rather more artistic than the idea actually was. The “colour” was essentially a label attached to a small amount of Bitcoin, allowing it to be treated as representing something else. A particular amount of Bitcoin could, for example, be associated with a share or another asset.
The important step was conceptual. Bitcoin no longer had to be thought of solely as a system for moving bitcoins around. Its transaction record could also carry information about something else. There were limitations. Bitcoin had not been designed as a general-purpose platform for creating arbitrary digital assets, and the techniques used by Colored Coins had practical constraints.
Still, the idea was out there. A blockchain could be used to say more than simply, “this person has some bitcoin.”
A blockchain for names
Then there was Namecoin.
Launched in 2011, Namecoin was based on Bitcoin's code but was designed for a different purpose: a decentralised system for registering names, including domains ending in .bit. A domain name has an interesting property in this context. You cannot have two identical .bit names occupying the same place in the registry. The name itself has to be unique. That meant Namecoin was already maintaining a public record of unique digital objects.
Nobody needed to call them NFTs. They weren't being sold in online marketplaces, and there was no standard called ERC-721. But some of the basic ingredients of digital ownership were beginning to appear. Then, in 2014, someone tried something much closer to what we now recognise as a digital artwork NFT.
Quantum
In May 2014, digital artist Kevin McCoy and technology entrepreneur Anil Dash were working together at the Seven on Seven event in New York. Their experiment became Quantum, a short, looping animation created by McCoy's wife, Jennifer. McCoy registered the work on the Namecoin blockchain. During the presentation, Dash bought it from him for $4.
The price is almost comically small by later NFT standards. At the time, however, nobody was trying to create a new asset class worth billions of dollars. They were demonstrating an idea. McCoy and Dash called their system Monegraph. It linked a blockchain record to a digital artwork, creating a way to identify and transfer ownership of the work. There was an awkward technical compromise hiding inside the demonstration. The artwork itself was not stored on the blockchain. The blockchain record pointed to it. That distinction would become a recurring problem in digital collectibles: owning a token and preserving the digital object associated with it are not necessarily the same thing.
Quantum is widely described as the first NFT. But “first NFT” depends partly on how the term is defined. Earlier blockchain projects had already created unique or scarce digital assets, and later historians have made different claims about where the NFT story properly begins.
The useful thing about Quantum is not the argument over the label. It is that the experiment made the connection between a specific digital artwork and a transferable blockchain record unusually explicit. That was a new idea.
Counterparty: making Bitcoin do more
Bitcoin itself was deliberately conservative. Changing its underlying rules was difficult, and that made it a frustrating place to experiment with more elaborate digital assets.
So developers found another route. In early 2014, Counterparty was created as a layer on top of Bitcoin. Rather than replacing Bitcoin, it used Bitcoin transactions to carry additional information. Counterparty could create its own assets and included a decentralised exchange where those assets could be traded.
Suddenly, Bitcoin was being used for things that looked rather different from ordinary currency. Artists could issue tokens. Game developers could create digital cards. Collectors could trade them. And because Counterparty ultimately relied on Bitcoin, the ownership records were secured by the Bitcoin blockchain itself.
This is where the early history starts to become rather eclectic.
One token, or one of many?
One interesting experiment came from Dutch artist and programmer JP Janssen. In 2014 he created OLGA, a one-of-a-kind Counterparty asset associated with a personal dedication. Around the same time, he also experimented with JPGOLD, an asset with a fixed supply of 1,000. These projects exposed an important distinction that would become central to NFT history.
A token does not necessarily have to be unique to be scarce. If I create 1,000 identical tokens, there are still 1,000 of them. They are limited in supply, but one unit is interchangeable with another. Janssen's experiments with fixed supplies helped explore the idea of limited-edition digital assets, a model that later projects such as Spells of Genesis and Rare Pepes would use. The vocabulary we use today can make this look more settled than it actually was.
At the time, people were experimenting with the economics and behaviour of digital tokens without necessarily having today's neat distinction between “fungible token” and “NFT.”
Digital trading cards arrive
In 2015, EverdreamSoft, led by Shaban Shaame, began developing Spells of Genesis. The project combined a mobile game with blockchain-based collectible cards. Early cards were issued on Counterparty, with the first card, FDCARD, released as part of a fundraising and awareness effort connected with Folding@home.
This was different from simply putting a piece of artwork on a blockchain. The token was becoming part of a game. The card was something you could collect, trade and use within a digital environment. The blockchain record was beginning to describe an object that had a purpose inside software rather than merely pointing to an image.
It was an early glimpse of an idea that would become commonplace years later: digital things could have lives of their own outside the database of the company that created them.
Pepe enters the blockchain
Then came one of the stranger chapters in the story. In 2016, artists began turning Pepe the Frog memes into blockchain collectibles on Counterparty. The resulting Rare Pepe project developed into a curated collection of digital trading cards. The humour is hard to miss. A cartoon frog that had originated as an internet meme was being transformed into scarce blockchain objects that people could collect and trade. The project also showed something about digital culture. These were not simply digital versions of traditional paintings. The subject matter itself came from internet culture.
Artists were making cards for an audience that already understood memes, online communities and cryptocurrency. The blockchain became another place where internet culture could be collected, traded and preserved. The first Rare Pepe cards appeared in 2016, before Ethereum-based NFT projects became widespread. The collection eventually grew to many series and hundreds of cards.
It is difficult to imagine a more internet-native form of collecting.
Ethereum changes the toolbox
By 2015, another blockchain had arrived. Ethereum was designed to do something Bitcoin was deliberately not trying to do: run programs on the blockchain. These programs became known as smart contracts.
The term sounds grander than it needs to. A smart contract is essentially code stored and executed on the blockchain. Instead of merely recording that money moved from A to B, the blockchain could run rules written by a developer. That opened up more complicated possibilities.
One early example was Etheria, a virtual world built around a grid of hexagonal land tiles. Its public release was deployed in October 2015, only a few months after Ethereum itself launched. Players could purchase tiles, name them and trade them. Ownership of those tiles was handled directly by the Ethereum contract, years before ERC-721 existed.
There was something slightly odd about Etheria's timing. It was creating persistent, player-owned digital land before there was a recognised standard for NFTs. The technology existed; the terminology and infrastructure were still catching up.
Another early Ethereum experiment, Terra Nullius, allowed users to claim pieces of digital land and attach permanent messages to them. The pattern was becoming clearer. People were no longer merely asking whether a blockchain could record ownership. They were experimenting with what owning a digital object might actually mean.
Curio Cards and digital art
In May 2017, Curio Cards appeared on Ethereum. Created by Thomas Hunt, Travis Uhrig and Rhett Creighton, the project presented itself as an online gallery of digital art. The cards featured work from multiple artists and were released as a collection rather than as a single artist's series.
Curio Cards is particularly interesting because the project was becoming more sophisticated. It was not simply a picture with a token attached. Its creators were thinking about artists, editions, provenance and the storage of the artwork itself. The contract used IPFS for the artwork and metadata, an early attempt to avoid relying entirely on a conventional web server. At the time, this was still a niche experiment. There was no NFT boom. There was no expectation that a complete Curio Cards collection would one day appear at Christie's. The people buying the cards in 2017 were participating in something whose later importance was not yet obvious.
The punks and the cats
A few weeks after Curio Cards, CryptoPunks appeared. Larva Labs created 10,000 pixel-art characters, each with its own combination of traits, and initially made them available for free to people with Ethereum wallets. The idea was simple enough to explain in a sentence: here were 10,000 characters, and each one was represented by its own blockchain token. But the technology behind these early projects was still being worked out.
Then came CryptoKitties. In late 2017, users could collect, trade and breed digital cats. Each cat had its own characteristics and could produce new cats with combinations of traits inherited from its parents. This time, the experiment reached well beyond the small cryptocurrency community. CryptoKitties became popular enough that its transactions contributed to severe congestion on Ethereum. The blockchain had suddenly become a place where people were breeding cats.
It was an unusually clear demonstration of a problem Ethereum developers now faced. If every project invented its own way of describing and transferring unique digital objects, wallets and applications had to understand each one separately. There had to be a common language.
ERC-721
That common language became ERC-721. The standard was formally proposed in January 2018 by William Entriken, Dieter Shirley, Jacob Evans and Nastassia Sachs. It defined a common interface for non-fungible tokens on Ethereum. The important part is the token ID.
An ERC-721 contract can associate each individual token with a unique tokenId. The combination of the contract address and token ID identifies a particular token. One token can therefore be distinguished from another, even when both belong to the same collection. That may sound like a small technical detail. Imagine a box containing 500 identical prints. If the prints are simply interchangeable units, the blockchain can record that someone owns 37 of them. It does not necessarily care which 37. With individually identified tokens, the blockchain can instead record that you own token 137. That distinction made it much easier for wallets, marketplaces and other applications to work with unique digital objects in a consistent way.
There is an interesting wrinkle here. ERC-721 did not invent the idea of individually identifiable digital objects. CryptoKitties, CryptoPunks, Etheria and earlier projects had already been exploring it. The standard gave developers a shared framework for doing it.
The experiments came first. The standard followed.
The technology leaves the laboratory
By the end of the 2010s, many of the pieces were in place. There were blockchains capable of recording ownership. There were systems for issuing scarce digital assets. Artists had experimented with tokenised artwork. Games had experimented with player-owned objects and virtual land. Collectors had built communities around trading cards and internet memes. And Ethereum now had a common standard for individually identifiable tokens.
NFT marketplaces made these objects easier to discover and trade. Artists could sell digital work directly to collectors. Game developers could experiment with portable assets. Collectors could build collections whose ownership was recorded publicly rather than existing only inside a company's database.
The idea also spread beyond Ethereum. Other blockchains developed their own approaches to digital collectibles, often emphasising lower fees, different technical designs or different communities. Looking back, it is tempting to draw a straight line from Hal Finney's 1993 crypto trading cards to today's NFT marketplaces.
The real story is less tidy. There are several possible starting points, depending on what exactly we mean by an NFT. There are abandoned experiments, competing technical approaches and projects created years before “NFT” became the standard name for what they were doing. That is what makes the history worth following.
The early digital collectibles were not all prototypes of today's NFTs. They were people trying things. Sometimes they were solving a technical problem. Sometimes they were trying to make digital art collectible. Sometimes they were building a game. And sometimes they were simply curious about what would happen if you put a picture, a name, a trading card or even a cartoon frog onto a blockchain.
Only later did these experiments begin to look like pieces of the same story. The history of NFTs is not really the story of someone inventing a finished technology. It is the story of people gradually working out what a digital object could be, and what it might mean to own one.